Sector

Social media 

Challenge

Wizz is a social media platform aimed at users aged 13-17, strictly adhering to each country’s digital consent age. Ensuring users experience a safe environment where they feel comfortable expressing themselves freely is paramount. To enhance user safety, Wizz needed a reliable way to estimate and verify user ages, placing users into appropriate age groups — specifically, with users of the same age, one year younger or one year older — to significantly reduce inappropriate interactions, especially between minors and adults. To maintain integrity and security, users cannot change their age after account creation. 

Approach

Wizz uses a layered age assurance process. Users self-declare their date of birth during onboarding. This is then validated using Yoti's facial age estimation technology, including anti-spoofing checks. If the estimated age matches closely enough with the declared age, account creation continues. Sumsub serves as a fallback verification method if Yoti encounters technical issues or unclear estimations. 

Wizz chose Yoti for their strong privacy standards, accuracy, competitive pricing and market leadership in age estimation. Sumsub was selected due to their quality, ease of implementation, responsive support and seamless in-app experience. 

Impact 

The age assurance strategy significantly improved user safety and platform integrity. For example: 

• Users are accurately grouped into appropriate age categories.  

• The strategy created a safer, more inclusive environment, allowing users to comfortably engage and express themselves, as confirmed by user surveys. 

• It supports effective moderation aligned with user ages. 

• It significantly reduces fake profiles and scammers. 

One ongoing challenge is deciding how to manage users who fail age verification. 

Future plans 

Wizz performs benchmarking of the best age estimation tools on the market every six months. Wizz continuously works to improve age estimation accuracy by collaborating with Yoti to refine models for challenging demographics (age, gender, skin tones) and exploring alternative verification methods to support users unable to pass standard checks, aiming for greater inclusivity without compromising safety.